Summary.

The objective of the study was to use D-optimal strategy in bioequivalence study design to search optimal time points and to estimate their ability to permit reconstruction of reference plasma concentrations/time relationship by test D-optimal population pharmacokinetic (PK) model with Bayes prediction. There were used test and reference products ampicillin concentrations from 14 healthy volunteers. D-optimal time points were defined at 1st, 4th and 6th hour. With concentrations at these time points were developed D-optimal population PK models of test and reference products. Parametrization was performed via absorption rate constant, KA (1/h), volume of distribution normalized to the body weight, VS1 (L/kg), and systemic clearance normalized to the body weight, CLS1 (L/h/kg). There were not observed significant differences between mean estimates of these model parameters of test and reference population PK models. Reference concentrations (at 1st, 4th and 6th h) were predicted by Doptimal test population PK model. Correlation coefficient between predicted versus observed concentrations was r = 0.874, p < 0.001. The test model predicted reference concentrations with statistic significance, which presumes that two drug formulations were bioequivalent.